Governance, Approval Authority and Conflicts of Interest in Business Origination

Investment analysis begins with the chain of decisions

Three questions need separate answers: does a business opportunity exist, does the proposed product fit that opportunity, and can the institution accept and manage the resulting risk? A project may have plausible demand without having existing receivables. A company may generate cash but have unresolved identity or registration issues. A transaction may fit an individual approval limit while creating excessive concentration in one economic group. Each question requires its own evidence. A persuasive presentation cannot establish demand, product suitability and institutional capacity simultaneously, even when the underlying business appears promising.

Investors should reconstruct the decision chain before interpreting growth indicators. Who introduced the proposal? Who checked the information? Who approved the risk, reviewed the agreement and authorized disbursement? Who checks whether the money returns? The answers help distinguish a repeatable process from dependence on personal interventions with little documentation. Judgment remains necessary, but it should be attributable, reasoned and open to challenge. An investor should be able to identify both the decision and the evidence available when it was made.

Origination, correspondent services and financing have distinct boundaries

Origination is a broad term. It may cover promotion, identifying businesses, organizing information and referring proposals. Acting as a correspondent for an institution requires an appropriate contractual arrangement, a defined scope and compliance with applicable requirements. A technology company does not become a correspondent simply by collecting documents or integrating systems. The agreement should identify permitted activities, communication channels, training, supervision, customer service and limits on representation. These boundaries help prevent a business owner from receiving promises that the institution has never authorized.

The financing institution retains the responsibilities assigned to it by regulation. Hiring a channel does not reduce supervision to commercial trust or automatically make a third party's assessment a substitute for institutional controls. Investors should examine how the institution monitors the contractor, receives evidence, corrects deviations and stops inappropriate activities. A contract can allocate tasks and consequences between parties, but that allocation does not, by itself, remove responsibilities toward customers or authorities. The relevant question is whether the institution can demonstrate how it exercised the responsibilities that remain with it.

KEZS connects entrepreneurial projects and potential funders through technology and intermediation. Its project-oriented method can help organize intended uses of funds, documents, assumptions and diligence questions. That potential usefulness does not establish authorization to perform regulated activities or the existence of a correspondent appointment. Investors conduct their own diligence and make their own decisions; KEZS does not assign a credit score. Evaluating any channel requires preserving the distinction between organizing an opportunity and exercising an institutional responsibility. References to a method should not be read as evidence that a particular regulated arrangement is already in place.

A credit policy is an economic choice that can be checked

A useful credit policy starts by describing the risks an institution is willing to carry. It should cover target businesses, financing purposes, permitted instruments, repayment sources, concentration, security, documentation and monitoring conditions. It also needs practical exclusions. Examples include an unexplained source of funds or a structure that presents hoped-for sales as existing receivables. Without exclusions that actually affect decisions, policy can become a justification written after the commercial choice has already been made. Stating an appetite for risk is meaningful only when it changes which proposals proceed.

Each rule should connect a reason with an evidence requirement. If exposure to one buyer is limited, the policy must explain how related companies are identified and consolidated exposure is calculated. If proof of delivery is required, it should specify acceptable documents, who checks their consistency and when additional confirmation is needed. Phrases such as 'sufficient documentation' offer little control unless they identify the facts to be demonstrated and the person responsible for accepting the evidence. A file can contain many documents while leaving the decisive question unanswered.

For investors, the crucial comparison is between the stated policy and the portfolio actually originated. A sample should include approved, rejected, exceptional and renegotiated transactions. If almost every proposal needs an exception, the product may be poorly designed, the policy outdated or commercial pressure excessive. Each explanation calls for a different response: redesign the product, formally revise risk appetite or restrict origination. Policy changes should govern future decisions and preserve the historical record. Rewriting a rule to make an earlier breach disappear destroys the control's value and obscures what decision-makers knew at the time.

Approval authority must aggregate exposures and constrain powers

Approval authority combines the power to decide, a limit and conditions for using that power. A sound matrix considers amount, risk, term, concentration, document quality and the nature of any exception. Approving a small transaction with regular documentation is different from waiving an essential check. Some unresolved issues should block progress until clarified, rather than becoming monetary exceptions that a director can approve because the amount is low. Authority to accept commercial risk does not include authority to disregard applicable obligations. The matrix should make that distinction visible.

Consider a wholly hypothetical matrix allowing a committee to approve exposure of up to R$ 500,000 per economic group where no material exception exists. A higher level must decide above that amount. If a group already owes R$ 180,000 and requests another R$ 400,000, total exposure would reach R$ 580,000. Splitting the request into two proposals of R$ 200,000 does not change the result. The process must aggregate contracts, undisbursed commitments and known relationships according to policy criteria. It should also retain the calculation available at approval, so later reviewers can reconstruct the decision.

One person should not independently control customer records, approval and disbursement. Smaller teams can combine dual authorization, independent review and restricted access, while considering cost and actual staffing capacity. Powers of attorney, temporary substitutes and emergency access also need limits. An organizational chart tells investors less than the record of a real transaction: who decided, under which permissions, against what total exposure, and with what explanation for departures from policy. A named reviewer is useful only if that person has enough information and authority to question the decision before the funds move.

Conflicts emerge through compensation and relationships

A conflict of interest arises when someone can influence a decision and benefit from an outcome that differs from the interests they are expected to protect. An originator paid only on disbursement has an incentive to close transactions. An analyst measured mainly on speed may ask fewer questions. An executive with an ownership interest in a supplier may favor that supplier. None of these circumstances, on its own, proves fraud or a poor decision. The governance problem is leaving the incentive hidden while allowing its beneficiary to control the assessment without effective counterweights.

A conflict map should cover direct compensation, paid referrals, ownership interests, relevant personal relationships, related-party transactions and benefits tied to product selection. Each relationship requires a response: disclosure may suffice in some cases; others require withdrawal from the decision, independent analysis or refusal to enter the arrangement. A generic annual declaration does not replace an update when a new relationship emerges. Nor does an investor's acknowledgment make every practice acceptable. The assessment must consider what influence the person has, what information they control and whether the proposed safeguard can work.

Channel commissions may combine a payment on contracting with later criteria involving document quality, substantiated complaints or operational performance. Each design can have adverse effects. Excessive withholding may raise channel costs and encourage selection of only straightforward customers. Rewards for low delinquency may encourage concealment of renegotiations. Complementary indicators and clear challenge procedures are therefore necessary. Investors need to know who pays, which event triggers payment, how the amount is calculated and whether the business, institution or vehicle bears the cost. Disclosure of the commission should make its economic effect understandable, including when it is incorporated into the price.

KYC must explain the business and the people controlling it

Know Your Customer, or KYC, for a business involves more than confirming a registration number. The institution needs to understand the activity, representatives, signing powers, control structure and ultimate beneficial owners under applicable requirements. It should also assess whether the purpose, amounts and expected transactions fit the business. Funding equipment purchases raises different questions from advancing receivables owed by a small number of buyers. Information gathering should be proportionate to risk and support an explanation of the conclusion. Collecting a standard bundle without understanding the transaction leaves important gaps.

Distinguish a document received, a document validated and a fact corroborated. Corporate documents may identify a manager without establishing authority for a specific transaction. An invoice may show that it was issued without proving delivery, acceptance, absence of dispute or absence of an earlier assignment. A bank account entered on a form needs ownership and purpose checks before disbursement. Records should preserve these distinctions so that an 'attachment received' checkbox cannot be mistaken for completed analysis. Verification should address the particular fact on which the financing decision depends.

In preventing unlawful activity, warning signs require treatment consistent with their nature and the institution's duties, including appropriate analysis and reporting by responsible personnel. Unusual urgency, resistance to identifying controllers or unexplained instructions to pay third parties justify further checks; they are not automatic findings of criminal conduct. Investors should examine independence, escalation capacity and confidentiality safeguards. Asking for explanations does not authorize disclosure of restricted assessments or improper notification to a customer about protected procedures. The process must allow a concern to be examined without turning suspicion into an unsupported public allegation.

The instrument determines the rights and controls required

Product governance begins with a direct question: what does the investor or funder actually acquire? A loan creates payment obligations under agreed terms. An assignment of receivables requires analysis of the claims and the legal effects of their transfer. Equity carries the rights and risks of ownership. A convertible instrument depends on its conversion and payment provisions. A donation to a nonprofit is neither an equity investment nor a promise of financial return. Presenting these instruments as interchangeable versions of one investment hides differences in enforceability, repayment, participation and potential loss.

The instrument must also fit the business's legal structure. A Brazilian MEI, or individual microentrepreneur, cannot admit equity partners while remaining a MEI. An equity proposal therefore requires appropriate treatment of the business structure; completing a form does not resolve the incompatibility. Similarly, expected sales do not become existing receivables eligible for assignment because they appear in a forecast. A policy may permit lending supported by expected cash generation, but it must recognize that exposure for what it is and apply suitable analysis and terms. Different evidence supports an existing claim and a future commercial expectation.

Public distribution of securities and other regulated financial activities require appropriate structures and providers. Technology, a private agreement or a CNAE economic activity classification is not general permission to conduct them. Materials from Banco Central, CVM, Portal do Empreendedor and ANPD are entry points for examining the Brazilian framework, not proof of a platform's authorization. Before considering economic attractiveness, investors should identify the instrument, issuer or debtor, service providers, rights, costs and risks. The label used in a presentation cannot replace this analysis, and an attractive business purpose does not establish that the proposed legal structure is suitable.

A numerical example: from proposal to approval

Suppose a fictional business presents R$ 600,000 in receivables for an advance. Every amount and condition in this example is hypothetical arithmetic, representing neither market pricing nor an existing transaction. Analysis excludes R$ 60,000 associated with disputed deliveries and R$ 40,000 whose possible previous assignment remains unresolved. This leaves R$ 500,000 treated as eligible in the example. Applying a hypothetical advance limit of 80% produces approved principal of R$ 400,000. The remaining R$ 100,000 is the initial difference between eligible receivables and principal, rather than an additional payment made by any party.

That difference is not cash deposited in a reserve, a recovery guarantee or assured profit. It is simply an initial economic cushion within the chosen structure. The hypothetical agreement charges an R$ 8,000 fee deducted at disbursement and R$ 12,000 in total financing remuneration collected from subsequent receipts. The principal obligation therefore begins at R$ 400,000, while the business initially receives R$ 392,000 net. The R$ 12,000 remuneration is accounted for separately from principal repayment, and the fee is not charged again. Keeping these categories separate prevents confusion between the amount owed, cash delivered and financing cost.

Assume the business has no other known exposure under the aggregation criteria and no material exception applies. The transaction fits the hypothetical R$ 500,000 approval limit discussed earlier. Approval should record the R$ 100,000 excluded, the evidence supporting the R$ 500,000 eligible balance, debtors, maturities and conditions for release. If another R$ 180,000 of related exposure is discovered before disbursement, total exposure becomes R$ 580,000 and requires a new decision at the competent level. The earlier approval cannot justify disregarding new information. Eligibility analysis and authority checks must both remain valid when funds are released.

The cash schedule reveals the risk of releasing surpluses

In the same example, the R$ 500,000 of eligible receivables falls due in three installments: R$ 200,000 on day 30, R$ 150,000 on day 60 and R$ 150,000 on day 90 after disbursement. These are relative intervals, not calendar dates. For simplicity, the hypothetical agreement schedules principal repayments of R$ 160,000, R$ 120,000 and R$ 120,000, together with remuneration of R$ 6,000, R$ 4,000 and R$ 2,000 respectively. These amounts are stipulated for the exercise; they are not derived from an advertised interest rate. Their purpose is to make the movement of cash traceable.

If each installment is received in full into the contractually designated collection account, allocation proceeds as follows. On day 30, R$ 166,000 pays principal and remuneration, and R$ 34,000 is released to the business. On day 60, R$ 124,000 meets those obligations, and R$ 26,000 is released. On day 90, R$ 122,000 meets the final obligations, and R$ 28,000 is released. The amounts reconcile: R$ 400,000 of principal, R$ 12,000 of remuneration and R$ 88,000 of surpluses total R$ 500,000. Each release transfers cash out of the structure at a particular point in the repayment schedule.

The business consequently receives R$ 392,000 initially and R$ 88,000 during collection, totaling R$ 480,000 in the full-payment scenario. The R$ 20,000 difference from R$ 500,000 consists of the R$ 8,000 fee and R$ 12,000 remuneration, excluding other costs or taxes. This calculation does not measure an investor's net return: expenses, losses, funding arrangements and participation terms may change that outcome. The schedule does show who receives each amount and when a release reduces the protection still available. A correct overall total does not remove the need to examine the sequence.

Delay, dispute and loss require different scenarios

Keep the full collections on days 30 and 60, including releases of R$ 34,000 and R$ 26,000. Now suppose the final R$ 150,000 installment produces only R$ 90,000 because a commercial dispute causes a loss of R$ 60,000. On day 90, R$ 120,000 of principal and R$ 2,000 of remuneration are due. The shortfall for that stage is R$ 32,000. Allocation of the available R$ 90,000 between principal and remuneration follows the contractual priority, but that allocation does not change the total cash deficiency. The shortfall exists regardless of which category is recorded as unpaid first.

Cumulative collections are R$ 440,000, exceeding the scheduled R$ 412,000 of principal and remuneration. However, R$ 60,000 has already been released to the business. The funder receives R$ 166,000 plus R$ 124,000 plus R$ 90,000, totaling R$ 380,000. This explains why the initial cushion does not automatically protect the exposure through final maturity. Recovering the missing R$ 32,000 depends on recourse rights, security, solvency and actual recovery costs. The numerical difference between receivables and principal cannot establish whether an enforceable and collectible claim exists against another party.

An alternative structure could retain intermediate surpluses in a contractual reserve until release conditions are met. In that alternative, R$ 440,000 collected could meet the scheduled R$ 412,000 and leave R$ 28,000 before additional costs. The business would receive less cash during the period. Retention does not eliminate legal or operational risks. A simple delay requires a different assessment: estimate the revised receipt timing, carrying cost and source of liquidity. Treating delay as definitive loss, or a real dispute as routine lateness, produces inappropriate decisions and communications. Reserve design therefore involves a trade-off between ongoing business liquidity and protection against later shortfalls.

Settlement and reconciliation complete the accountability chain

Settlement is the financial execution that fulfills an obligation under the applicable terms. Reconciliation compares records to establish which obligation each movement relates to. Receiving R$ 200,000 is insufficient to close the first installment in the example. The institution must identify the payer, receivable, receipt timing, account, any deductions and allocation. A bank balance matching the expected total may conceal a duplicate payment, money from another business or a receivable that remains incorrectly open. Matching the total is one check; establishing the identity and treatment of each payment is another.

The routine should compare the agreement and its schedule with the account statement, receivables records and accounting entries. Differences need classification, an owner and a resolution deadline. An unidentified payment remains pending until clarified. A manual adjustment requires justification and review. A refund should return to the appropriate source after validation. An institution should not use an unknown deposit to make portfolio delinquency look better or offset obligations of different customers for operational convenience. These practices can obscure the true position even when the cash balance itself is accurate.

The architecture must also identify account ownership, authority to move funds, payment providers and the destination of money if operations are interrupted. Specific controls depend on the product and legal structure; calling an account 'segregated' does not establish protection from competing claims. Investors should seek evidence of completed reconciliations, old differences, manual adjustments and resolution times. Systems may automate matching, but ambiguous cases need recorded human decisions. Automation that quickly clears the wrong receivable can make an error harder to detect. Its usefulness depends on reliable matching rules, appropriate access and a process for unresolved items.

Operational roles and evidence must outlast individual staff

Clear responsibilities prevent a proposal from advancing on assumptions. Commercial staff present the product within authorized boundaries and record its source. Customer onboarding checks identity, representation and documents. Risk assesses repayment capacity, structure, concentration and conditions. Legal reviews the relevant instruments, while compliance handles matters within its remit. Operations checks disbursement conditions; treasury executes authorized movements; reconciliation and collections track fulfillment. The exact arrangement varies, but every material task needs an owner. Handoffs should make clear what has been verified and what remains outstanding, rather than passing an apparently complete file onward.

In a small institution, one person may perform compatible roles. The assessment should focus on where someone could both create and conceal an error. A person changing bank details should not independently authorize the following payment. Someone negotiating a document waiver should not be the sole person confirming its acceptability. Where advance separation is impractical, a compensating review needs to occur while it can still help, by someone with access and authority. Reviewing an unrecoverable disbursement months later offers limited protection. The cost of additional review should be weighed against the exposure it is intended to control.

Evidence should allow reconstruction without relying on an executive's memory. It includes the policy version, information analyzed, sources, questions, responses, decision and fulfillment of conditions. New documents should not silently replace those used at approval. The history helps distinguish a reasonable decision with a bad outcome from a decision made without support. That distinction matters to investors: losses are part of risk, while an inability to explain how exposure was accepted weakens confidence in the portfolio as a whole. A complete record supports learning and accountability without implying that a documented decision was necessarily correct.

Disagree and negotiate without making risk disappear

Imagine commercial staff arguing that the R$ 60,000 of disputed deliveries should be included because the buyer 'always pays.' Risk asks for delivery confirmation and resolution of the dispute. The disagreement should become a verifiable question: what fact would make the receivable eligible, and what exposure remains if the dispute continues? Payment history may be relevant, but it does not prove acceptance of this delivery. The decision record should preserve the disagreement and outcome, including a rejection of the commercial request. A useful committee record explains the basis of the decision rather than merely recording that discussion occurred.

Negotiation may change the amount, term, security, reserve, reporting obligation or disbursement condition. Every change redistributes costs and risks. Reducing principal may leave the project underfunded. Requiring a larger reserve may make the transaction impractical for the business. Additional security may increase expense without providing meaningful recovery. Alternatives should be assessed with the customer and the competent decision-makers, without presenting a contractual amendment as an automatic solution. If no structure fits the policy, rejection is a valid outcome. A higher price cannot by itself resolve a missing right, unreliable evidence or an incompatible product.

Commercial urgency can also conflict with the time needed for analysis. A controlled response may be conditional approval that prevents release until missing evidence is supplied, provided the approach is permitted and clearly communicated. Signing the agreement must not cause the condition to disappear. Investors should check whether the institution tracks conditionally approved proposals that never satisfy requirements, allows technical challenges and protects staff who identify inconsistencies. A committee without a real opportunity to disagree may record only apparent consensus. Conditional approval remains useful only if someone verifies fulfillment before disbursement and retains evidence of that check.

Data and technology need explicit boundaries

A project-oriented process can bring intended use of funds, budgets, contracts, responsible parties and execution assumptions into a coherent file. Its usefulness lies in comparison and inquiry. Does the amount requested match the demonstrated need? Does delivery depend on a related supplier? Does projected cash flow include the interval before collection? Records should distinguish verified facts, applicant statements and assumptions. Organization can improve diligence, but it neither turns a forecast into revenue nor independently validates its source. A well-presented budget still requires examination of the evidence and dependencies behind its figures.

Data sharing requires a defined purpose, an appropriate legal basis, access controls and retention consistent with applicable obligations. Investor reporting may use aggregation or restricted access according to need, protecting personal data and business secrets. Demonstrating that a control occurred does not require distributing complete documents to everyone. Exports, copies and supplier access also need controls. When a commercial contract ends, permissions should be reviewed and information still subject to retention obligations handled appropriately. The question is which information each party needs to perform its role, with enough evidence to support review and without unnecessary disclosure.

If blockchain is used, record integrity does not establish factual truth, legal title, reserve solvency or payment certainty. Private documents should not be placed on public blockchains. Even a technical reference needs assessment of what it could reveal and who could link it to individuals or businesses. Automated classification and artificial intelligence similarly require validation, supervision and limits on use. These possibilities do not assert any existing KEZS functionality or replace the investor's own decision. Technology can preserve a false statement accurately, so the reliability of the original evidence remains a separate question from the integrity of its record.

Indicators, failures and corrections must tell a consistent story

A governance dashboard should follow the proposal through to cash collection. Useful indicators include documents returned by channel, approvals with exceptions, group concentration, information changed after approval, arrears by aging band, commercial disputes and unresolved reconciliation differences. Analysis by channel and by groups of transactions originated in comparable periods helps distinguish origination quality from portfolio maturity. Comparing a newly formed portfolio with one whose obligations have already fallen due can create a misleading impression of performance. The observation period and exposure being measured need to be clear before differences are interpreted.

Rates and denominators must remain consistent. If renegotiated transactions leave an arrears indicator, reporting should separately show their volume and subsequent behavior. Recoveries should not erase the original loss event; cancellations and repurchases need identification. The example's contractual R$ 12,000 remuneration cannot be presented as received before collection. Likewise, the R$ 88,000 surplus returned to the business is not investor profitability. Operational and accounting records must support every classification. Otherwise, a change in presentation may appear to improve performance without changing the underlying cash or risk.

Suppose an unauthorized bank-account change is detected before release. Correction requires suspending payment, confirming instructions through an independent channel, preserving history and reviewing access. If money has already left, recovery measures and incident handling become necessary, with communication to affected parties according to need and applicable duties. Both situations require investigation of the cause: excessive permissions, missing validation or false instructions. Merely correcting the account record leaves the mechanism available for another incident. Investors should examine whether remediation addresses that mechanism, who owns the corrective action and what evidence shows it was completed.

An investment conclusion should state its limits

Diligence can conclude with a memorandum connecting the investment rationale, evidence and conditions. Instead of stating that governance is 'robust,' an investor might record that approval samples were checked, group exposure aggregation was observed and dual disbursement authorization was identified, but sufficient evidence on old reconciliation differences was unavailable. A conclusion is more useful when it states what was observed, what remains unresolved and how the gap affects the decision. Missing evidence does not prove a failure, but it limits justified confidence. Specific observations also make subsequent monitoring more meaningful.

Before accepting exposure, material questions about the channel arrangement, responsibilities, approval limits, conflicts, instrument rights, costs and movement of money should be resolved. Negotiable points can become documented conditions with an owner and evidence of fulfillment. Essential unanswered questions can justify postponement or rejection. The same discipline applies during monitoring: a change of control, increasing concentration, more exceptions or recurring failures may require review even while payments remain punctual. Timely payments do not establish that every other control remains effective, particularly when the portfolio or operating structure changes.

The practical objective is to follow an entire transaction from its first introduction to its final reconciled movement and explain where loss could occur. In the numerical example, the key finding was that early surplus releases changed the protection available at final maturity. In another structure, the decisive issue might be signing authority, a commercial conflict or an unsuitable instrument. Governance provides ways to identify and address these matters responsibly. It offers no guarantee of approval, repayment, profitability, liquidity or preservation of capital. Its contribution is a more explainable decision and a clearer account of the uncertainty being accepted.

Frequently asked questions.

References and context

Institutional reference points for checking applicable rules and guidance. Calculations and scenarios are original examples, not quotes, historical results or individualized recommendations.

• Banco Central — correspondentes no País ↗
• CVM — Resolução 88, texto e alterações ↗
• CVM — Resolução 175, fundos de investimento ↗
• Portal do Empreendedor — perguntas sobre MEI ↗
• ANPD — orientações e perguntas frequentes ↗

AI-assisted educational content. Decisions require verification of the information, instruments and providers in each actual case.